PRECISION DAIRY FARMING: POTENTIALS, PITFALLS, AND CRYSTAL BALL GAZING Jeffrey Bewley, PhD, PAS 2012 OABP Fall Continuing Education Meeting

Where I Come From

Kentucky Dairy Industry 85,000 dairy cows across 810 dairy farms

Technological Marvels Tremendous technological progress in dairy farming (i.e. genetics, nutrition, reproduction, disease control, cow comfort) Modern dairy farms have been described as “technological marvels” (Philpot, 2003) The next “technological marvel” in the dairy industry may be in Precision Dairy Farming

Changing Dairy Landscape Fewer, larger dairy operations Narrow profit margins Increased feed and labor costs Cows are managed by fewer skilled workers

Consumer Focus Continuous quality assurance “Natural” or “organic” foods Greenhouse gas reductions Zoonotic disease transmission Reducing the use of medical treatments Increased emphasis on animal well-being

Information Era Unlimited on-farm data storage Faster computers allow for more sophisticated on-farm data mining Technologies adopted in larger industries have applications in smaller industries

Cow Challenges 1. 1.Finding cows in heat 2. 2.Finding and treating lame cows 3. 3.Finding and treating cows with mastitis 4. 4.Catching sick cows in early lactation 5. 5.Understanding nutritional status of cows a. a.Feed intake b. b.Body condition (fat or thin) c. c.Rumen health (pH/rumination time)

Fatness or Thinness Mobility Hoof Health Mastitis Respiration Rumination/pH Temperature Milk content Heart rate Animal position/location Chewing activity Lying/ standing behavior Methane emissions Feed intake Areas to Monitor a Dairy Cow Manure and Urine

Can technologies provide us the answers we’ve been looking for?

Precision Dairy Farming Using technologies to measure physiological, behavioral, and production indicators Supplement the observational activities of skilled herdspersons Focus on health and performance at the cow level

Precision Dairy Farming Make more timely and informed decisions Minimize medication (namely antibiotics) through preventive health Optimize economic, social, and environmental farm performance

Precision Dairy Farming Benefits Improved animal health and well-being Increased efficiency Reduced costs Improved product quality Minimized adverse environmental impacts Risk analysis and risk management More objective (less observer bias and influence)

UK Herdsman Office

Ideal PDF Technology Explains an underlying biological process Can be translated to a meaningful action Low-cost Flexible, robust, reliable Information readily available to farmer Farmer involved as a co-developer at all stages of development Commercial demonstrations Continuous improvement and feedback loops

Mastitis Detection Exciting Precision Dairy Farming applications May increase likelihood of bacteriological cure May reduce duration of pain associated with mastitis (animal well-being) May reduce the likelihood of transmission of mastitis between cows May prevent the infection from becoming chronic Potential to separate abnormal milk automatically Brandt et al., 2010; Hogeveen et al., 2011

Mastitis Detection Challenges Meeting sensitivity (80%) and specificity (99%) goals (Rasmussen, 2004) Calibration across time Automatic diversion or alert? Recommended action when an alert occurs with no clinical signs Multivariate systems provide best results Costs (both fixed and variable)

Electrical Conductivity Ion concentration of milk changes, increasing electrical conductivity Inexpensive and simple equipment Wide range of sensitivity and specificity reported Affected by sample time, milk viscosity, temperature, and sensor calibration Results improve with quarter level sensors Improved results with recent algorithms Most useful when combined with other metrics Brandt et al., 2010; Hogeveen et al., 2011

Milk Color Color variation (red, blue, and green) sensors in some automatic milking systems Reddish color indicates blood (Ordolff, 2003) Clinical mastitis may change color patterns for three colors (red, green and blue) Specificity may be limited

Temperature Not all cases of mastitis result in a temperature response Best location to collect temperature? Noise from other physiological impacts

Thermography May be limited because not all cases of mastitis result in a temperature response Difficulties in collecting images Hovinen et al., 2008; Schutz, 2009 Before Infection After Infection

Automated CMT or WMT CellSense (New Zealand) Correlation with Fossomatic SCC 0.76 (Kamphuis et al., 2008) Using fuzzy logic, success rates (22 to 32%) and false alerts (1.2 to 2.1 per 1000 milkings), when combined with EC were reasonable (Kamphuis et al., 2008) Costs?

< 200, ,000 — 400, ,000 — 800, ,000 — 2,000,000 > 2,000,000 SCC value Works like a traffic light

Mastiline Uses ATP luminescence as an indicator of the number of somatic cells Consists of 2 components In-line sampling and detection system, designed for easy connection to the milk hose below the milking claw Cassette containing the reagents for measuring cell counts

Spectroscopy Visible, near-infrared, mid-infrared, or radio frequency Indirect identification through changes in milk composition AfiLab uses near infrared – –Fat, protein, lactose, SCC, and MUN May be more useful for detecting high SCC cows than quantifying actual SCC

Biosensors and Chemical Sensors Biological components (enzymes, antibodies, or microorganism) Enzyme, L-Lactate dehydrogenase (LDH), is released because of the immune response and changes in cellular membrane chemistry Chemical sensors: changes in chloride, potassium, and sodium ions, volatile metabolites resulting from mastitis, haptoglobin, and hemoglobin (Hogeveen, 2011) Brandt et al., 2010; Hogeveen et al., 2011

Milk measurements Progesterone – –Heat detection – –Pregnancy detection LDH enzyme – –Early mastitis detection BHBA – –Indicator of subclinical ketosis Urea – –Protein status

4Sight-Fionn Technologies Northern Ireland Photosensitive optic beams across barns Software recognizes cows ID’s when cows cross beams

Estrus Detection Efforts in the US have increased dramatically in the last 2 years Producer experiences are positive Changing the way we breed cows Only catches cows in heat Real economic impact GEA Rescounter II AFI Pedometer + SCR HR Tag/AI24 DairyMaster MooMonitor/ SelectDetect LegendBouMatic HeatSeeker II

SCR HR Tag Measures rumination timeMeasures rumination time Time between cud bolusesTime between cud boluses Monitor metabolic statusMonitor metabolic status

Rumen pH Illness Feeding/drinking behavior Acidosis

Vel’Phone Calving Detection

CowManager Sensoor TemperatureTemperature ActivityActivity RuminationRumination Feeding TimeFeeding Time

Alanya Animal Health Behavioral changes Temperature Lying/Standing Time Grazing Time Lameness Estrus Detection (multiple metrics) Locomotion Scoring

Rumination, Drinking, Eating Behavior Lying, Standing, Steps RumiWatch

Greenfeed measures methane (CH 4 ) Select for cows that are more environmentally friendly Monitor impacts of farm changes (rations) on greenhouse gas emissions

StepMetrix Lameness detection BouMatic

Real Time Location Systems Using Real Time Location System (RTLS) to track location of cows (similar to GPS) Better understand distribution of animals within barns Information used to design better barns and modify existing barns Behavior monitoring-implications for estrus detection, time at feedbunk, social interactions Black et al.

GEA CowView Feeding time Waiting time Resting time Mounting Distance Covered

MonitorParameter Measured 3-D acceleration/movementBehavior ElectromyogramMuscle activity Skin potentialVegetative-nervous reaction Skin resistanceVegetative-emotional reaction Skin temperature/Environmental temperature Thermoregulation

Economic Considerations Need to do investment analysis Not one size fits all Economic benefits observed quickest for heat detection/reproduction If you don’t do anything with the information, it was useless Systems that measure multiple parameters make most sense Systems with low fixed costs work best for small farms

Purdue/Kentucky Investment Model Investment decisions for PDF technologies Flexible, partial-budget, farm-specific Simulates dairy for 10 years Includes hundreds of random values Measures benefits from improvements in productivity, animal health, and reproduction Models both biology and economics

Net Present Value (NPV) Simulation Results Results from 1000 simulations Positive NPV=“go” decision/make investment

Tornado Diagram for Deterministic Factors Affecting NPV NPV establishes what the value of future earnings from a project is in today's money.

From Purdue to Poor Due Did I get the wrong PhD?

From Purdue to Poor Due Did I get the wrong PhD?

Sociological Factors Labor savings and potential quality of life improvements affect investment decisions (Cantin, 2008) Insufficient market research Farmers overwhelmed by too many options (Banhazi and Black, 2009) – –Which technology should I adopt? – –End up adopting those that are interesting or where they have an expertise – –Not necessarily the most profitable ones

Australian Case Study R&D tends to focus on the device rather than the management system within which the device will be used “Return on investment is only achieved through subsequent improvement in the farming system and it is here that people are key” Not enough focus on farmer adaptation and learning Need more formal and informal user networks Eastwood, 2008

Why Have Adoption Rates Been Slow?

Reason #1. Not familiar with technologies that are available (N =101, 55%)

Reason #2. Undesirable cost to benefit ratio (N =77, 42%)

Reason #3. Too much information provided without knowing what to do with it (N =66, 36%)

Reason #4. Not enough time to spend on technology (N =56, 30%)

Reason #5. Lack of perceived economic value (N =55, 30%)

Reason #6. Too Difficult or Complex to Use (N =53, 29%)

Reason #7. Poor technical support/training (N =52, 28%)

Reason #8. Better alternatives/easier to accomplish manually (N =43, 23%)

Reason #9. Failure in fitting with farmer patterns of work (N =40, 22%)

Reason #10. Fear of technology/computer illiteracy (N =39, 21%)

Reason #11. Not reliable or flexible enough (N =33, 18%)

Reason #99. Wrong College Degree (N =289, 100%)

Future Vision New era in dairy management Exciting technologies available and in development New ways of monitoring and improving animal health, well-being, and reproduction These analytic tools will be a source of competitive advantage Economics and people factors will determine adoption and success

Questions? Jeffrey Bewley, PhD, PAS 407 W.P. Garrigus Building Lexington, KY Office: Cell: Fax: